With the work reported in this manuscript we have essentially contributed to the electrical characterization and modelling of high voltage MOSFETs, more particularly DMOS architectures such as X-DMOS and L-DMOS able to sustain voltages ranging from 30V to 100V. The technology information and the investigated devices have been kindly provided by AMIS, Belgium (former Alcatel Microelectronics). In general, all the initial defined targets in term of the orientation of our work, as defined in the introduction chapter, have been maintained along the progress of the work. However, sometimes, based on the obtained results we have decided to pay more attention to some less explored topics such as the hot carrier impact of DMOS capacitances and the combined effect of stress and temperature, which initially were not among the planned activities. However, we believe that we have contributed to some of the planned targets. We experimentally validated the concept of intrinsic drain voltage; a modeling concept dedicated to the modeling of HV MOSFET and demonstrated its usefulness for the DC and AC modelling of HV devices. We proposed an original mathematical yet quasi-empirical formulation for the bias-dependent drift series resistance of DMOS transistor, which is very accurate for modelling all the regimes of operation of the high voltage device. We combined for the first time such a model with EKV low voltage MOSFET model developed at EPFL. We also have reported on models for the capacitances of high voltage devices at two levels: equivalent circuits for small signal operation based on VK-concept and large signal charge-based models. These models capture the main physical charge distribution in the device but they are less adapted for fast circuit simulation. In the field of device reliability, we have originally contributed to the investigation of hot carrier effects on DC and AC characteristics of DMOS transistors, with key emphasis on the degradation of transistor capacitances and the influence of the temperature. At our knowledge, our work reported in this chapter is among the first reports existing in this field. We have essentially shown that the monitoring of capacitance degradation if mandatory for a deep understanding of the degradation mechanisms and, in conjunction with DC parameter degradation, could offer correct insights for reliability issues. Even more, we have shown situations (by comparing two fundamental types of stresses) when the capacitance degradation method by HC is much more sensitive than DC parameter degradation method. Of course, some of the combined stress-temperature investigations were too complex to find very coherent explications for all the observed effects but our work stress out the interest and significance of such an approach for defining the SOA of high voltage devices, in general. Overall, our work can be considered as placed at the interface between electrical characterization and modelling of high voltage devices emerging from conventional low voltage CMOS technology, continuing the research tradition in the field established at the Electronics laboratory (LEG) of EPF Lausanne.